Diagnose Why Fewer Food Orders Are Being Handed Over to Couriers

You are the product manager of a food delivery app, and the count of fulfilled orders has been sliding downward. For this metric, an order counts as fulfilled at the moment the restaurant hands it to the delivery agent who will carry it to the customer. Walk through how you would find the root cause of the decline. ```hint Split the count first A falling number of handoffs can come from fewer orders entering the system or from a smaller share of them reaching a courier. Work out which is happening before you examine customers, restaurants or couriers. ``` ### Clarifying Questions - How large is the decline, over how many weeks has it built up, and did anything similar happen in the same season last year? - Are fewer orders being placed, or is a smaller share of placed orders reaching the handoff, for example because more orders are cancelled before pickup? - How is a handoff recorded: does the courier confirm pickup in their app, does the restaurant mark the order as collected, or both? - Is the decline spread across all cities and order types, and does the count include orders that customers collect themselves or that restaurants deliver with their own staff? ### What a Strong Answer Covers - Use of the stated definition: a handoff to a courier, not a delivery to the customer. - A check that the decline is real (event logging, late data, counting changes) before any product analysis. - A decomposition into orders placed and the share reaching handoff, then localization by segment and by step of the order flow. - Hypotheses across customers, restaurants and couriers, plus internal changes and external factors, each paired with the data that would confirm or rule it out. - An investigation order suited to a gradual trend, and how a suspected cause is confirmed before acting. ### Follow-up Questions - Orders placed are flat, but cancellations of orders that never get a courier have doubled during dinner hours in two large cities. What do you look at next? - The decline is concentrated among the restaurants that used to process the most orders. Which hypotheses move to the top, and how do you test them? - How would you set up monitoring so that a slide like this is caught within days instead of weeks?